Analysing Routes Generated by Planning-based Centralised Traffic Routing

Abstract

Centralised Traffic Routing offers a global perspective of the entire region, enabling effective route finding for vehicles that navigate through the region. Recently, automated planning techniques were leveraged for centralised traffic routing. However, centralised routing techniques might generate different routes for vehicles even if they share the same origin and destination and arrive at a similar time. Also, centralised routing techniques are usually computationally demanding. In this paper, we analyse the outputs of a planning-based centralised traffic routing method in terms of the number of different routes per journey, differences between traveled distance, and travel time among these (different) routes. Then, we propose a method that leverages the found routes and their relative distribution to vehicles for an online centralised routing.